VAITP

`python-multipart` is a streaming multipart parser for Python. When using form data, `python-multipart` uses a Regular Expression to parse the HTTP `Content-Type` header, including options. An attacker could send a custom-made `Content-Type` option that is very difficult for the RegEx to process, consuming CPU resources and stalling indefinitely (minutes or more) while holding the main event loop. This means that process can't handle any more requests, leading to regular expression denial of service. This vulnerability has been patched in version 0.0.7.

sentry-sdk is the official Python SDK for Sentry.io. A bug in Sentry's Python SDK < 2.8.0 allows the environment variables to be passed to subprocesses despite the `env={}` setting. In Python's `subprocess` calls, all environment variables are passed to subprocesses by default. However, if you specifically do not want them to be passed to subprocesses, you may use `env` argument in `subprocess` calls. Due to the bug in Sentry SDK, with the Stdlib integration enabled (which is enabled by default), this expectation is not fulfilled, and all environment variables are being passed to subprocesses instead. The issue has been patched in pull request #3251 and is included in sentry-sdk==2.8.0. We strongly recommend upgrading to the latest SDK version. However, if it's not possible, and if passing environment variables to child processes poses a security risk for you, you can disable all default integrations.

A vulnerability has been found in the CPython `venv` module and CLI where path names provided when creating a virtual environment were not quoted properly, allowing the creator to inject commands into virtual environment "activation" scripts (ie "source venv/bin/activate"). This means that attacker-controlled virtual environments are able to run commands when the virtual environment is activated. Virtual environments which are not created by an attacker or which aren't activated before being used (ie "./venv/bin/python") are not affected.

A defect was discovered in the Python “ssl” module where there is a memory race condition with the ssl.SSLContext methods “cert_store_stats()” and “get_ca_certs()”. The race condition can be triggered if the methods are called at the same time as certificates are loaded into the SSLContext, such as during the TLS handshake with a certificate directory configured. This issue is fixed in CPython 3.10.14, 3.11.9, 3.12.3, and 3.13.0a5.

llama-cpp-python is the Python bindings for llama.cpp. `llama-cpp-python` depends on class `Llama` in `llama.py` to load `.gguf` llama.cpp or Latency Machine Learning Models. The `__init__` constructor built in the `Llama` takes several parameters to configure the loading and running of the model. Other than `NUMA, LoRa settings`, `loading tokenizers,` and `hardware settings`, `__init__` also loads the `chat template` from targeted `.gguf` 's Metadata and furtherly parses it to `llama_chat_format.Jinja2ChatFormatter.to_chat_handler()` to construct the `self.chat_handler` for this model. Nevertheless, `Jinja2ChatFormatter` parse the `chat template` within the Metadate with sandbox-less `jinja2.Environment`, which is furthermore rendered in `__call__` to construct the `prompt` of interaction. This allows `jinja2` Server Side Template Injection which leads to remote code execution by a carefully constructed payload.

python-jose through 3.3.0 allows attackers to cause a denial of service (resource consumption) during a decode via a crafted JSON Web Encryption (JWE) token with a high compression ratio, aka a "JWT bomb." This is similar to CVE-2024-21319.

The `mjml` PyPI package, found at the `FelixSchwarz/mjml-python` GitHub repo, is an unofficial Python port of MJML, a markup language created by Mailjet. All users of `FelixSchwarz/mjml-python` who insert untrusted data into mjml templates unless that data is checked in a very strict manner. User input like `<script>` would be rendered as `<script>` in the final HTML output. The attacker must be able to control some data which is later injected in an mjml template which is then send out as email to other users. The attacker could control contents of email messages sent through the platform. The problem has been fixed in version 0.11.0 of this library. Versions before 0.10.0 are not affected by this security issue. As a workaround, ensure that potentially untrusted user input does not contain any sequences which could be rendered as HTML.

python-jose through 3.3.0 has algorithm confusion with OpenSSH ECDSA keys and other key formats. This is similar to CVE-2022-29217.

The “socket” module provides a pure-Python fallback to the socket.socketpair() function for platforms that don’t support AF_UNIX, such as Windows. This pure-Python implementation uses AF_INET or AF_INET6 to create a local connected pair of sockets. The connection between the two sockets was not verified before passing the two sockets back to the user, which leaves the server socket vulnerable to a connection race from a malicious local peer. Platforms that support AF_UNIX such as Linux and macOS are not affected by this vulnerability. Versions prior to CPython 3.5 are not affected due to the vulnerable API not being included.

The Snowflake Connector for Python provides an interface for developing Python applications that can connect to Snowflake and perform all standard operations. Prior to version 3.12.3, when the logging level was set by the user to DEBUG, the Connector could have logged Duo passcodes (when specified via the `passcode` parameter) and Azure SAS tokens. Additionally, the SecretDetector logging formatter, if enabled, contained bugs which caused it to not fully redact JWT tokens and certain private key formats. Snowflake released version 3.12.3 of the Snowflake Connector for Python, which fixes the issue. In addition to upgrading, users should review their logs for any potentially sensitive information that may have been captured.